413 Croonian Lecture :\#151 ; On Globulins .
By W. B. Hardy , M.A. , F.R.S. ( Being the Croonian Lecture delivered May 25 , 1905.\#151 ; MS .
Received April 9 , 1907 .
) Globulins are a class of proteids which occur in both animal and vegetable tissues .
They are peculiar in the complexity of their relations to electrolytes .
Insoluble in water , they are soluble in low concentrations of acids , alkalies , or neutral salts .
In presence of acids the globulin is electro-positive , in presence of alkalies it is electro-negative , in presence of neutral salts it is electrically neutral .
Electrically active globulin ( i.e. , dissolved by acids or alkalies ) is precipitated by minute amounts of neutral salts ; also , no matter what its electrical state may be , or how dissolved , globulins are precipitated by neutral salts near the saturation point of the latter .
The problem I propose to consider is their diversified relation to electrolytes .
Connected with this problem is another , namely , the relation of solutions of globulins to colloidal solution .
Do they form hydrosols at all , and , if so , to what extent ?
Krafft urged the colloidal nature of soap solutions , because , within certain limits of temperature and concentration , they gelatinise , and have the vapour pressure of pure water .
Kahlenberg and Schreiner , however , regard soap solutions as being crystalloid in character , because .over the whole range of concentration the soap is a good electrolyte\#151 ; it ionises and undergoes hydrolytic splitting , like other salts of a weak acid and strong base .
Smits , again , by measurement of the vapour pressure over a wide range of concentration , is convinced that above a critical concentration the soap passes wholly into the colloidal state .
Exactly the same points arise in connection with proteid .
Waymouth Reid has shown that the proteids in a solution exert no measurable influence upon the vapour pressure .
On the other hand , solutions of globulins are relatively good conductors over the whole range of concentration through which it is possible to follow them .
The dilution curve shows no break indicating a general change of state , eveu when concentration is pushed to the point where fluidity almost vanishes .
Again , colloids , as a class , are chemically inert .
But globulins react actively with acids or alkalies to neutralise them .
Globulin solutions , even -at extreme concentration , form syrups and not true gels , except , perhaps , under special circumstances .
Clearly , therefore , it is pertinent to ask whether globulins form colloidal solutions .
VOL. lxxix.\#151 ; 'b .
2 h 414 Mr. W. B. Hardy .
[ Apr. 9 , The globulin used throughout this research is that which is precipitated from ox serum by dilution and slight acidification .
Solution by Acids or Alkalies .
Electro-positive or electro-negative globulin .
When acid or alkali is gradually added to a suspension of globulin in water , , the opaque suspension gradually changes to a clear transparent fluid .
Between the opaque suspension , from which the globulin settles on standing , , to the stage of limpid transparency there is no break .
The most minute addition of acid or alkali ( in the absence of salts ) converts the suspension into a non-settling " solution " of low grade , which is opaque white from the large size of the particles of globulin dispersed throughout it .
Further addition of acid or alkali raises the grade of solution .
The globulin particles become smaller and smaller until complete transparency is reached .
The process can be reversed by dialysis , save that , at the lowest grade attainable , no precipitate settles in the absence of disturbing factors .
With the first addition of acid or alkali , and at the last stage of dialysis , , the globulin is electrically active , so that it moves in an electric field , the direction of movement being that which one would expect if it combined with the acid or the alkali to form a salt , and the specific velocity is quite , or nearly , at its maximum value .
A rise in the grade of solution is , on the whole , associated with a fall in the specific velocity , never in my experience with a rise in this value .
The electric charge on the surface of the globulin particles therefore appears to reach maximal density when the solutions are still of exceedingly low grade .
These facts can best be explained by assuming that true salt formation occurs .
This agrees with the view of all previous workers on the interaction of proteids with acids or alkalies ( Sjoqvist , Bugarsky and Liebermann , and Cohnheim ) .
Since globulins combine either with acids or bases , they have both an acid and a basic function\#151 ; they are amphoteric electrolytes .
The globulin salts ionise in solution ; therefore , in an electric field the entire mass of proteid moves .
They also hydrolyse , but the hydrolysis offers special features resembling those which Jordis has pointed out in the case of sodium silicate , and Chevreul in the case of soaps .
In both of these cases hyper-acid salts are formed , while hyper-acid salts are formed on dialysis of a solution of globulin by alkali , and hyper-basic salts on dialysis of a solution of globulin by acid .
By dialysis the degree of hyper-acidity or hyper-basicity can be raised , , but , as Jordis finds in the case of silica , with continuously increasing difficulty .
With the rise in the degree of hyper-acidity or hyper-basicity , the " grade " of solution diminishes , but , in the case of globulins , precipitation .
1907 .
] On Globulins .
415- does not occur .
Electrically active solutions of globulin cannot be precipitated by dialysis , nor at any stage do they cease to be electrically active .
If G be used to denote globulin , the equation of hydrolysis or dialysis would be : * GHS + 3^ HOH = ( GHOH)y ( GHS)*-y+y HS , or x GB+y HOH = ( GH)y ( GB)^\#151 ; y+yBOH .
In dialysis , the ratio x/ y varies continuously .
It is as indeterminate as the ratio between the combining salts has been found to be by Bodlander , Abegg , Sherrill , and others in the double salts of mercury and silver .
Clearly a relation of this kind agrees with van Bemmelen 's definition of absorption compounds as chemical combination with variable composition .
Kemoval of water from a solution of acid or alkali globulin does not produce precipitation , and the dried gummy residue reabsorbs water and passes slowly again into a state of solution .
Acid and alkali globulin , therefore , form , with water , solutions which have the feature characteristic of colloidal solutions , in that there are no saturation points .
In order to compare the solvent power of different acids or alkalies , it is necessary to fix upon some arbitrary point , such as the point of minimal opalescence .
By the use of a system of controls and proper illumination this point furnishes very concordant values .
Measured in this way , it appears that for strong and medium acid\#174 ; solvent power is measured by the number of gramme molecules present , not by the number of gramme equivalents : HC1 = H2S04 = H3PO4 .
Very weak acids have a lower solvent power : HC1 = 5HA = \#177 ; 30,000H3BoO3 .
These relations are explained by the very weak basic function of globulin .
Salts with weak acids are much hydrolysed , and to reach the same grade of solution an excess of acid is needed in order to lower the degree of hydrolysis .
With alkalies , the weak alkali NH4OH dissolves as well as the strong alkalies , owing to the fact that globulin acts as an acid of considerable strength .
The acid and basic functions were measured by the well-known methods\#151 ; the catalysis of cane sugar and of methyl acetate\#151 ; and the acid function found to be much the greater .
The molecular relation noticed above recalls the salts of amido acids : RCOOH+HC1 -pNHaHCl ^COOH \#187 ; \#166 ; dNH3 ^COOH + H2S04 p .nh2h2so4 ^COOH 2 h 2 Mr. W. B. Hardy .
[ Apr. 9 , Data derived from measurements of electric conductivity , and from the behaviour with indicators , support the view that true salts are formed , and that the globulin acts much more strongly as an acid than as a base .
Direct Measurement of the Specific Velocity of Globulin ."\#151 ; This was carried out by the boundary method .
As the basic function of a globulin is weaker than its acid function , the salts GHS ( globulin + acid ) will be hydrolysed more than the salts GB ( globulin + base ) .
Therefore , in the equation of hydrolysis given above , will be larger for the former than the latter .
Similarly , comparing salts with HA and with HC1 , yfx will be greater for the former .
How with increase in the value of y/ x the size of the proteid particles increases also .
By dialysis , the size can be increased until the particles diffract white light .
The growth in the case of solutions of sodium silicate can be traced by the appearance of molecular states incapable of passing through a parchment membrane .
In a similar way , by continuous addition of AgCl to a solution of NaCl , molecular states are produced large enough to diffract white light .
In each of these cases these large molecules are ionic , that is to say , they take part in the electric transport .
And their specific velocity is exceptionally high ; that of ( AgCl)* Cly being 57 x 10-5 at 18 ' , whereas Ag at infinite dilution is only 58 .
" Ions " of this order of magnitude have the properties of matter in mass .
Each is defined by a surface and moves under the influence of a surface contact difference of potential .
With their appearance the fluid ceases to be homogeneous .
It has internal surfaces .
I propose to call such ions colloidal ions , or pseudoions .
Their specific velocity is high , and , within wide limits , is independent of their size , and is controlled by the laws of electrical endosmose .
These conclusions are borne out by the boundary measurements and by the electrical conductivity of colloidal solutions .
For instance : theoretically the proportion of pseudoions in the following solutions of globulin should be:\#151 ; In HA \gt ; in HC1 , \gt ; in NaOH , and the specific velocities at 18 ' were found to be 23 x 10-6 , 11'5 x 10 \gt ; 7'6 x 10-5 .
By Ostwald 's law of the relation of specific ionic velocity to the number of atoms in an ion , the value 23 x 10-5 is an utterly impossible one for an ordinary ion of the magnitude of the proteid molecule .
Solution by Neutral Salts.\#151 ; Globulins are dissolved by neutral salts owing to the formation of molecular compounds ( G.BS ) .
These compounds aie readily decomposed by water with liberation of insoluble globulin 1907 .
] On Globul .
( G.BS + HOH = G.HOH + BS ) .
Therefore they are stable only in presence of a large excess of salt .
Hence the solvent power of salts is from 200 to 500 times less than that of acids or alkalies .
Hence , also , the presence of the ' globulin lowers the electric conductivity of the salt to only a small extent ( 2 to 6 per cent. ) .
I A double salt of the form AB .
CD .
ionises according to AB .
CD ^ AB .
C + D. | I have never succeeded in detecting any trace of such ionisation in the case of globulin and neutral salts ; possibly owing to the extreme instability of the ions .
The proteid does not move in an electric field , and it seems to take no part in the electric transport .
\ Owing to the insolubility of the dissociation products , the globulin can be precipitated from its solution with neutral salts by simple dilution .
No degree of dilution will precipitate globulin from solution in dilute acid or alkali . !
The compounds of globulin and alkalies ( GB ) are more readily dissolved by neutral salts than is simple globulin .
Compounds of globulins and acids are insoluble by neutral salts , being decomposed with liberation of the acid .
If on the analogy of an amido acid a globulin combines owing to the presence of the NH2 and the COOH group , then it forms salts with alkalies by replacement of hydrogen , and with acids by the change of the trivalent .
nitrogen of the amido group to the pentavalent form .
H2 h2 .I n i. n +NaOH = R\lt ; ^ n + HOH .
X'\lt ; OH '\lt ; ONa ii .
H2 m r/ N a +HC1 X'\lt ; OH H2 ' I H Cl , O OH Where , then , does the neutral salt link on ?
There are two possible places\#151 ; the unsatisfied valencies of the nitrogen , or the unsatisfied valencies of the upper O of the COOH group .
According to the oxonium theory , O has a maximum valency O^s , but 0= is much less stable than Nl ; therefore , one would expect the linkage to be by the amide group\#151 ; iii .
H2 bSN n +NaCl Nc\lt ; oh ' ' .Na .
= b/ *\lt ; gi .
Nc\lt ; ' ^OH Mr. W. B. Hardy .
[ Apr. 9 , And this would account for the fact that the acid can be turned out of the combination with globulin by an excess of neutral salt , but the alkali cannot be .
It is easy to prove by the boundary method that the globulin , when dissolved by salts , takes no part in the electric transport .
Dissociation of the salt-globulin compound by the water may be regarded as being practically completely suppressed by the excess salt present .
This being the case , the diminution of the electric conductivity of the salt by the globulin may be used as a measure of the fraction of the salt actually combined with it .
In the following table a few values are given .
Comparison is made between solutions containing the same amount of salt per litre ; but in the one case the solution is saturated with globulin , in the other case it is a simple solution of the salt in water .
No correction is made for the diminution of the molecular conductivity of the salt owing to replacement of a portion of the water by globulin , since it would amount to less than O'Ol per cent. No. Ill is of the nature of a control , to prove that electrolytic impurities adherent to the purified globulin may be neglected .
The salt solution was made up with a dialysate in equilibrium with the suspension of globulin .
Serum centrifuged , diluted to 10 vols .
, globulin precipitated by HA .
The precipitate suspended in a large volume of distilled water and collected by the centrifuge , and re-suspended in water freed from gases by boiling .
Specific conductivity of the water employed , 6 x 10 6 rec .
ohm , 18 ' .
I. Ox globulin suspended in water .
100 c.c. = 4*08 grammes dry globulin .
Temperature , 18':\#151 ; Specific Specific conductivity , conductivity , salt globulin , salt water , K salt globulin Salt .
Normality .
x 10s .
x 10\#174 ; .
K salt water K2S04 ... ... . .
0 *1716 1445 1539 0 *938 " ... ... 0 -1679 1411 1506 0 -937 NaCl ... ... . .
0*27 1850 1950 0*949 Specific resistance of the globulin suspension 3*1 x 10 ~5 .
II .
Sheep globulin .
100 c.c. = 4*4 grammes :\#151 ; NaOl ... ... .
, .
I n .oftcQ 2145 2280 0*938 " ... ..1 ' 2150 2278 0*943 The measurements in II were duplicated in order to test the degree of accuracy of the measure ' ments of volume .
III .
Ox globulin dialysed until in equilibrium with water saturated with toluol :\#151 ; MgS04 ... ... 0*11 517 543* 0*95 In this case the salt solution ( * ) was made up with the dialysate , not with water .
[ Note.-\#151 ; In a recent paper* it is stated that globulin in solution does not alter the * Mellanby , ' Journ. of Physiology , ' vol. 33 , p. 354 , 1905 .
On Globulins .
1907 .
] conductivity of the salt used to dissolve it .
The conductivity of the salt " is the same as that of a similar solution in pure water.5 ' This statement is directly contrary to my own observations .
It leads me to add a few words as to the precautions which are necessary in dealing with proteid solutions .
Electrodes coated in the ordinary way with platinum black are quite useless .
The proteid in their immediate neighbourhood is changed very rapidly and very drastically .
An instance selected at random will show the magnitude of the effect .
18 ' .
Globulin dissolved by dilute NaOH .
3'3 per cent , globulin .
Readings at intervals of two minutes approximately :\#151 ; 2560 ohms .
2542 .
" 2537 " 2530 " 2522 " 2538 " The rise in the last observation followed upon slight shaking of the cell to wash the changed portion of the solution away from the electrodes .
Platinum grey electrodes , prepared as Whetham directs , give no trouble .
Readings with globulin solutions of all kinds remain quite constant after being for 48 hours and longer in contact with them .
Even with platinum grey electrodes , however , it is necessary so to adjust the cell that the observed resistance does not fall under about 500 ohms .
" With a lower resistance , and therefore a larger current , the readings become unsteady and too low .
] On the assumption that the whole of the drop in electric conductivity is due to association of the salt with the globulin , the above figures show that , at the particular concentration and temperature , 1 gramme of dry globulin combines with 33 x 10-5 gramme equivalent of NaCI and 26 x 10-5 gramme equivalent of K2SO4 , while to dissolve 1 gramme of dry globulin 10 x 10-5 equivalent of alkali , or 18 to 36 equivalents of acid are needed .
The figures for salt and acid are of the same order of magnitude , and this is what might be expected if salt and acid combine in the same way with the proteid .
Appendix .
( Added December , 1906 .
) Viscosity of Solutions of Globulins.\#151 ; The measurements of this value , very briefly touched upon in the lecture , have been amplified .
They show the following interesting features :\#151 ; In dilute solutions of globulins the viscosity , as measured by the rate of flow through a capillary tube , is of the same order of magnitude for each of the three types\#151 ; acid , alkali , and salt globulin .
" With increasing concentration the viscosity increases , but the increase is much greater for alkaline globulin than for acid globulin , and for acid globulin I than for salt globulin .
The difference is very striking ; thus , at a concentraMr .
W. B. Hardy .
[ Apr. 9r\gt ; tion of 7*59 grammes globulin per litre , the transpiration times are in the ratio\#151 ; Water ... ... ... ... .
1 MgSC\gt ; 4 globulin ... .
4*66 HC1 " ... . .
15-5 NaOH " ... ... 679 In solutions of salt globulin , the globulin particles do not carry electricity " ionic " proteid is completely absent , therefore the higher viscosity of acid and alkaline globulin solutions is connected with the presence of " ionic"1 proteid , that is , of very large molecules carrying a charge .
Since globulin acts much more strongly as an acid than as a base , 'there will be in similar solutions of HC1 globulin and NaOH globulin a greater concentration of " ionic " proteid in the latter .
Here , again , increased concentration of " ionic " proteid goes with increased viscosity .
These conclusions agree with the observations of Reyher* on the viscosity of solutions of soap , and of Sackiirj* on solutions of the proteid caseinogen. .
It would , however , be rash to fix upon the simple proteid or soap ion as the agent .
The effect of dilution , and of varying proportions of acid or alkali , make it possible that the high viscosity is due to the presence of the molecular complexes , which carry a surface charge , and which I have called " colloidal " ions , or pseudoions .
Thus , further addition of small quantities of ammonia , beyond the amount necessary to dissolve the globulin to a transparent solution , decreases the viscosity .
Ammonia is a very weak base , which , , in solution , has a specific molecular conductivity much less than that of its-compound with globulin .
One cannot , therefore , refer the diminished viscosity to diminished ionisation of the ammonia-globulin compound , but the optical properties of the solution show that slight excess of ammonia does diminish the size of the particles of the globulin .
It raises the " grade " of the solution .
A comparison of the different types of solution of globulin raises , in a. special way , the question of the meaning to be attached to the phrase " colloidal solution .
" Mixtures of very diverse character have been classed as colloidal , some because they form jellies , others because they are physically heterogeneous , others because the osmotic pressure is exceedingly low , others , again , because the electric conductivity is abnormally low .
Of these , perhaps , the only constant attribute is physical heterogeneity .
Solutions which present any of the other distinctive features of the colloidal state always-contain particles which are large enough to scatter light .
* 'Zeits .
f. Physik .
Chem. , ' vol. 2 , p. 743 , 1888 .
t Ibid. , vol. 41 , p. 672 , 1903 .
1907 .
] On Globulins .
Amongst the attributes of colloidal solutions is the possession of an-abnormally high viscosity .
Judged in this respect , the salt globulin solutions-are only slightly colloidal as compared with the solutions of alkali or acid globulin .
They differ also from acid and alkali globulin in the relatively greater definiteness of the relations between the components .
Dilution of a solution of salt globulin , for instance , brings about a sharp separation of solid and liquid .
No such clean saturation points are found with the other two systems .
The definiteness of the .
relations makes it possible to compare the system\#151 ; neutral salt , water , globulin with other similar three-component systems .
Data sufficient for the purpose can be obtained from the measurement of the solubility of globulins in aqueous salt solutions made by Osborne and Harris , * and by Mellanby.f The curves which these authors give fail to-show the true relations , since only two co-ordinates are used , namely for the salt and the globulin ; the third component , water , is neglected .
For the following curves , trilinear co-ordinates are used in the manner recommended by Stokes .
Fig. 1 shows curves plotted to scale from values derived from Osborne and ) s Fio .
1 .
* ' American Journal of Physiology,5 vol. 14 , 1905 .
t Loc .
cit. Mr. W. B. Hardy .
[ Apr. 9 , Harris.* S is pure salt , W pure water , G pure globulin ( edestin ) .
The points on the curves , fixed by actual experiment , are shown in the usual way by dots or crosses .
A'B'C is the curve for Na2S04 , ABC the curve for NaCl .
The broken line extending to just beyond B is the probable form of the curve for H2S04 .
The diagram is drawn as an isotherm for NaCl , water , edestin , the areas for the other salts being left out .
The context is not very clear on the point , but the temperature of observation seems to have been 20 ' .
Each curve clearly consists of two distinct parts , AB , which is in .equilibrium with solid salt , and BC in equilibrium with globulin crystals .
The whole surface , therefore , consists of certain areas:\#151 ; ABC , all points within which represent a homogeneous state , namely , a fluid solution of globulin , salt , and water ; ASB , which is an area of heterogeneous states .
Each point in it represents a mixture which cannot exist as a uniform state , but divides into two phases , one a fluid solution on the curve AB , the other a solid which is pure salt .
WBG , again , is an area of mixture which cannot form a uniform state , but which divides into a fluid solution on CB , and into globulin crystals .
SBG , each point within which represents a mixture which divides into a fluid having the composition of B and solid salt and solid globulin .
In fig. 2 , portions of three isotherms for NaCl , water , serum globulin are Fig. 2 .
* Loc .
cit. 1907 .
] On Globulins .
plotted to scale , the data being derived from Mellanby 's paper .
They cover only a very small fraction of the whole curve , ABO , but they are sufficient to \#171 ; how , ( 1 ) that the curve encloses an area which is placed on the line WS , as in the preceding case .
This follows from the form of the curve , and from the fact that if sufficient salt be added , a point is reached on the line WS where .salt solution is in equilibrium with solid globulin .
( 2 ) That the area enclosed increases with a rise of temperature .
This last fact , together with some observations of my own , enable us to .approximate roughly to an isotherm for edestin at , say , 30 ' .
When edestin is present in sufficient quantity , on warming a further amount passes into .solution , so that we have a series of solutions in equilibrium with crystals .
Beyond a certain temperature the crystals fuse , and we then have two fluid layers in equilibrium , the lower layer ( as I have found by actual .analysis ) containing all three components .
Fig. 3 is a diagram of an isotherm at a higher temperature ; it shows the .area ABC enlarged , and it shows a new area , BOP , due to the appearance of S O P Fig. 3 .
; a second series of solutions ( the melted crystals ) .
Of the shape of this area we know nothing .
The curve ABC now contains three parts , AD in .equilibrium with S , DE with G- , and OF with the solutions which lie on OK .
Fig. 1 does not represent the conditions for blood globulin at any temperature , since , in presence of water alone , a solid solution of water in globulin is formed .
The the lines from the curve BC , therefore , instead \#171 ; of focussing at G , end on a curve of unknown slope , which starts from about Mr. W. B. Hardy .
[ Apr. 9 , .
the middle of the line WG .
In other words , the isotherms of blood globulins , , over the region where we have any information to guide us , are of the form ; shown in fig. 4 .
o p Fig. 4 .
This is as far as the known facts carry us .
It now remains to compare* the system with a similar one .
Such a one is found in water , sodium chloride , and succinic nitrile , which has been studied by Schreinemakers.* In both there is a pair of immiscible bodies , salt , and globulin , or salt and succinic nitrile , each of which is partially miscible with the third substance , water .
Fig. 4 , A and B are two isotherms reproduced from Schreinemakers ' ' paper .
It is obvious that , in all essential features , they resemble the isotherms for globulin-salt-water already given .
It is clear , from this comparison and from the low viscosity , that globulin-salt-water is a system showing few abnormal features .
It manifests decisive points of equilibrium , which resemble those in a comparable and purely crystalloid system .
The other two globulin systems are as decisively abnormal , especially in the high viscosity , and the absence of definite-transition points , and it is significant that , regarded from the purely physical standpoint , the essential difference should be the presence in the latter of large molecular aggregates , each carrying an electric double layer .
The presence of internal electrified surfaces adequately accounts for the-high degree of inertia which obscures the transition points .
What is the-source of the electrification ?
In all cases in which it can be traced , the-charge is due to molecular interaction of the type classed as chemical , and of that particular chemical type associated with the decomposition of neutral electricity , to which the name " ionisation " has been given .
Burtonf clearly shows that metals form hydrosols because they react with the solvent to form hydrides or hydroxides .
Acid globulin , alkaline globulin , and soaps are salts .
* ' Zeits .
f. Physik .
Chem.,5 vol. 23 , p. 417 , 1897 .
t ' Phil. Mag.,5 April , 1906 .
11907 .
] On Globulins .
-which form colloidal solutions because one of the radicles is of large size , is ^almost or quite immiscible with the solvent , and therefore readily forms molecular complexes which have the electric sign proper to the radicle when .dissociated from its fellow in the salt molecule .
Each colloid particle undoubtedly carries on its surface a charge , since , without exception , they move in a uniform electric field .
But in the system formed by each particle there cannot be any free electricity , any resultant .charge .
If there were , the system would , by Earnshaw 's theorem , be unstable , and settling would occur the more rapidly the greater the charge .
Experiment shows the opposite to be the case .
The charge on each particle , therefore , must be bound by an equal and opposite charge on the liquid face opposed to it .
The condition of double electric layers is therefore realised .
It has been shown by Waymouth Reid* that in a solution of proteid no .osmotic pressure can be traced to the proteid , and I find that when a .solution of globulin is dialysed against water until equilibrium is reached , the dialysee and dialysate have exactly the same electric conductivity .
These facts , added to others already stated , afford overwhelming evidence that in these solutions the chief portion of the proteid is , in effect , removed from solution by being gathered into masses of more than molecular dimensions which are separated from the rest of the solvent by a surface .
The masses .of probably hydrated proteid thus form an internal phase .
In colloidal .salts such as soaps , solutions of globulin in dilute acid or alkali , solutions of .caseinogen or of acid or alkali albumen , in which the internal phase is bounded by double electric layers formed by ionic interaction with the .external phase , the stability of the system with reference to forces such as gravity will depend , in part , upon the potential difference between the two faces round each particle .
The electric double layers will contribute to stability , since any movement of a particle through the fluid will develop free electricity in quantity proportional to the potential difference between the layers.-j* Salts in small amount diminish the stability of these colloidal systems and bring about concentration ; and salts , as has been shown experimentally by Wiedeman , Perrin , { Burton , S and others , diminish the potential difference in the condenser system round each particle .
The contribution which the electrification of the internal surfaces makes to mechanical stability is seen in the high viscosity of these solutions as * * * S * ' Journ. of Physiology , ' vol. 31 , p. 438 , 1904 . .
t Helmholtz , * Wied .
Ann. , ' vol. 7 , p. 337 , 1879 .
X'Journ .
de Chim .
Physique , ' vol. 2 , p. 61 , 1904 ; vol. 3 , p. 50 , 1905 .
S ' Phil. Mag. , ' November , 1906 .
On Globulins .
compared with solutions of globulin by salts which resemble them in containing large molecules , but differ in the total absence of electric layers from these molecules .
Solutions of globulins in salts introduce us to an entirely different aspect of colloidal solution .
Here , again , the proteid contributes nothing to the osmotic pressure , * and , therefore , must be withdrawn from true molecular solution ; but as all ionic interaction is suppressed by the large excess of salt present , there are no double electric layers formed , and in their absence viscosity is low and stable equilibrium conditions are rapidly reached .
The origin of the mechanical stability of this system , the nature of the influence which maintains the even dispersion of the proteid throughout the fluid , is difficult to explain , since any factor to which we can appeal seems also to involve the exertion of a finite osmotic pressure by the proteid .
If the distribution of the energy be considered , we arrive at the following\#151 ; Let the state of thorough mixing of the components be established , and let a movement under the influence of any mechanical force take place , so that the proteid becomes more concentrated in some region .
In this region , owing to the concentration of the colloidal salt , the degree of hydrolysis and of ionisation will be diminished , and the available energy thereby increased .
The equation of hydrolysis which Arrheniusf gives shows that hydrolysis and ionisation diminish faster than concentration increases ; therefore , with unequal distribution of the proteid the gain in available energy ( E\#151 ; ) will be greater in the regions of higher concentration than the loss in the regions of diminished concentration .
The system as a whole , therefore , would gain in available energy , and the change from uniform distribution to unequal distribution could not take place unless work were done on the system .
* " Waymouth Reid , loc. cit. t * Zeits .
f. Physik .
Chem. , ' vol. 5 , p. 16 , 1890 .

